Preliminary study of surgical audit for overall survival following gastric cancer resection.

ABSTRACT

BACKGROUND: Previous studies for surgical audit have focused on short-term outcomes, such as perioperative mortality. There has been no gold standard how to evaluate quality of care for long-term outcomes in surgical oncology. This preliminary study aims to propose a method for surgical audit targeting long-term outcome following gastrectomy for gastric cancer. METHODS: We prospectively investigated a set of variables relating to physiologic conditions, tumor characteristics and operations in patients who underwent gastrectomy for gastric cancer between June 2005 and July 2008 in 18 referral hospitals in Japan. Overall survival (OS) is the endpoint. Cox hazard regression analysis was used to generate a model to predict OS. The calibration and discrimination power of the model were assessed using the Hosmer-Lemeshow (H-L) test and area under the receiver-operating characteristic curve (AUC), respectively. The ratio of observed-to-estimated 5-year OS rates (OE ratio) was defined as a measure of quality. RESULTS: Among 762 patients analyzed, 697 (91%) completed the 5-year follow-up. The constructed model for OS exhibited a good discrimination power (AUC, 95% confidence interval 0.89, 0.86-0.91), which was significantly better than that for the UICC stage (0.81, 0.77-0.84). This model also demonstrated a good calibration power (H-L χ(2) = 27.2, df = 8, P = 0.77). The OE ratios among the participating hospitals revealed no significant variation between 0.74 and 1.1. CONCLUSIONS: The current study suggests the possibility of surgical audit for postoperative OS in gastric cancer. Further studies including high-volume centers will be necessary to validate this idea.